Chronic disability following stroke is a significant problem for Veterans that can affect a variety of daily activities. One area of importance in clinical assessment and treatment planning is the impact of stroke on driving safety. Unfortunately, formal assessments of driving fitness and related cognitive deficits (e.g., visuospatial changes) are often not conducted in left hemisphere stroke patients. The current project will address this gap by evaluating left hemisphere stroke patients on a state-of-the-art driving simulator, comparing their performance to that of both right hemisphere stroke patients and healthy controls. Driving variables assessed by the simulator include the number of collisions, unsafe lane crossings, speed exceedances, and reaction time to stop. Different types of driving errors will then be related to performance on a neuropsychological battery, which includes standardized and experimental measures of visuospatial ability, executive functioning, and language. Last, we will investigate the neural correlates of distinct types of driving errors using voxel-based lesion symptom mapping, which relates structural lesions to behavioral performance on a voxel-by-voxel basis. Participants for the current study include 40 left and 40 right hemisphere Veteran stroke patients with no prior neurologic or severe psychiatric history. A control group of 20 age- and education-matched Veterans with no neurologic or psychiatric history will provide additional context for interpretation of the data. Driving performance will be examined with a state-of-the-art driving simulator that has been demonstrated to have strong ecological validity and predictive validity with respect to on-road driving fitness. It is predicted that 60-70% of left hemisphere patients will exhibit impaired driving performance (“failed” or “needs training” rating), significantly higher than healthy controls but not significantly different from right hemisphere patients. It is expected, however, that left hemisphere patients will show a distinct pattern of performance from right hemisphere patients. For example, it is expected that LH patients will exhibit disproportionately more errors under complex driving scenarios (e.g., in a construction zone), whereas both stroke groups will exhibit significant visuospatial driving errors (unsafe lane changes and collisions) relative to controls. Partial least squares regression will then be used to identify which neuropsychological measures are most closely related to driving performance variables measured in the simulator. It is predicted that driving errors in left hemisphere patients will correlate with both visuospatial measures (e.g., Useful Field of View/Visual Search) and executive functioning measures (e.g., Trails B). We also propose to relate structural lesion data from high-resolution 3T MRI scans to different types of driving errors. This aspect of the project will utilize voxel-based lesion symptom mapping (VLSM) soft...